1. ^aSocial Sciences Methods and Data Institute, University of Nottingham, Nottingham NG7 2RD, UK. E-mail: ceesvandereijk@nottingham.ac.uk
2. ^bDepartment of Political and Social Science, European University Institute, 50016 San Domenico di Fiesole, Italy. E-mail: Mark.Franklin@EUI.eu
3. ^cInstitute for Migration and Ethnic Studies, University of Amsterdam, 1012 DL Amsterdam, The Netherlands
4. ^dDepartment of Communication Science, University of Amsterdam, 1012 CX Amsterdam, The Netherlands

¹Earlier versions were presented at the Annual Meeting of the Dutch Political Science Association ('Politicologendagen'), Noordwijk, 23–24 May 2002, and at the Workshop on 'Perceptions, Preferences and Rationalization: Overcoming the Problem of Causal Inference in the Study of Voting', Nuffield College, Oxford, 7–9 May 2004. We are indebted to the participants in those meetings for their comments and criticisms.

Abstract

Conventional wisdom holds that the state of the economy is closely linked to the outcomes of elections: incumbent governments tend to be rewarded for good economic times and punished for bad ones. It has been suggested that the 'subjective economy' — people's assessments of the state of the economy — links actual economic conditions to support for the government. Indicators of such assessments are therefore frequently included in voter surveys. Such subjective information has been criticized, however, as being endogenous: being caused by rather than a cause of vote choice. The purpose of this paper is to assess the relevance of the subjective economy for linking objective economic conditions and support for the government at the individual level. To this end, we specify and estimate a series of rivaling causal models, and compare them in terms of fit and parsimony. We find unambiguous support for the endogeneity interpretation of the subjective economy, which implies that it cannot play a sensible role in the causal elaboration of economic voting models.